Data Mining and Visualization Overview August 29, 2019 Data Mining - - PowerPoint PPT Presentation

Data Mining and Visualization – Overview

August 29, 2019

Data Mining and Visualization – Overview August 29, 2019 1 / 16

Motto In God we trust, all others bring data

attributed to William Edwards Deming∗ (1900-1993)

∗American statistician, known, in particular, for promotion of statistical methods in

industrial production and management

Data Mining and Visualization – Overview August 29, 2019 2 / 16

Organization

Instructor, Assistant, Webpage, etc.

Instructor: Krzysztof (Krys) Podg´

• rski

The easiest way to communicate is through e-mail: Krzysztof.Podgorski@stat.lu.se Webpage: https://krys.neocities.org Assistant: Johan Larsson, e-mail: Johan.Larsson@stat.lu.se The main source of the information about the course: webpage – visit it frequently as the material that is posted there will change. The web address is

https://krys.neocities.org/Teaching/DataMining/DataMining.html

Office hours each Monday between 16:00 and 17:00 or by appointment

Data Mining and Visualization – Overview August 29, 2019 3 / 16

Organization

Course Organization

Syllabus is available in the form of a the webpage (print it on your

• wn if a hard copy is more convenient for you).

Three parts of a big comprehensive “examination”: Assignments, Projects, Presentation Assignments – Individually at home plus discussion in the classroom Computer Projects – In groups of two: decide with whom would you like to work, mostly done in the computer lab. If not completed then it can be finished at home but then a printed report showing that the tasks have been completed have to be submitted. The same applies in the case of absence from a lab session. Presentation – 30min presentation of the own study based on a chosen data set (this will be discussed in detail when halfway through the course), a written rapport after the presentation needs to be submitted.

Data Mining and Visualization – Overview August 29, 2019 4 / 16

Organization

Grade

For each of the three parts score will be assigned on the scale from 0-100 This score contributes equally to the total score which is computed according to the formula: T = (S1 + S2 + S3)/3, where S1, S2, S3 represent scores for the corresponding parts. The final grade will be assigned according to the table: Percentage Grade 49 - 0 F 54 - 50 E 64 - 55 D 74 - 65 C 84 - 75 B 100 - 85 A

Data Mining and Visualization – Overview August 29, 2019 5 / 16

Assignments

Assignments – cover basics and main topics of the course

The total of five assignments, that are based on the covered course material although they do not cover the material completely. They will comprise of a set of simple questions that will help to clarify introduced topics. Some questions will also help in preparing to the lab sessions. Assignments will be worked out at home and due the week following the date they are posted in our schedule. They should be submitted in an electronic form (scanning of the text is allowed) through CANVAS. They will be available at the webpage but there can be last minute changes thus, please, download a copy only on the date at which they are listed in the schedule.

Data Mining and Visualization – Overview August 29, 2019 6 / 16

Assignments

Course contents and the textbook

An Introduction to Statistical Learning with Applications in R, by Gareth James, Daniela Witten, Trevor Hastie and Robert Tibshirani; 1st edition; Springer; 2013.

This is a simplified version of the book that is the most popular and most comprehensive monograph on the statistical data mining (see the next slide). According to the authors, and concurred by many, the material of the book could serve as basis for a sequence

• f courses equivalent to 30hp.

For us to adopt it within a 7.5hp course frame, significant reduction of the material is needed. For the same reason the course will not follow the ‘chronology’ of the textbook.

Cover 2.jpg

Data Mining and Visualization – Overview August 29, 2019 7 / 16

Assignments

The ‘other’ textbook

The Elements of Statistical Learning,

Data Mining, Inference, and Prediction

by Trevor Hastie, Robert Tibshirani, and Jerome Friedman; 2nd edition; Springer.

As mentioned, this is by far the most popular and most comprehensive monograph on the subject. Originally the course was designed with this book in mind. However, publication of its simplified version which is easier to follow within a regular course frame, shifted this monograph into the background of our presentation. Some of the final lectures are using the material that is

• nly available in this textbook thus it contains all the

material needed for the course.

Data Mining and Visualization – Overview August 29, 2019 8 / 16

Assignments

Access to the textbook

The authors of ‘An Introduction to Statistical Learning’ have created a webpage on which the book is accessible in the electronic format. Hard copy is also accessible at Springer. The mentioned webpage contains also other useful material such as programs, data, and erratum. The ‘other’ textbook, ‘The Elements of Statistical Learning’ is also available in the electronic form through Lund University’s online library due to the agreement of the university with Springer.

Data Mining and Visualization – Overview August 29, 2019 9 / 16

Assignments

Supplementary books

There is a plethora of other books published recently on Data Mining, Big Data, Data Analytics – a consequence of growing interests in the

• methodology. The following are examples (we may occasionally take a look

into some of them, especially when the final projects are discussed): Depending on your background, some of them can be more accessible than

• thers.

Data Mining and Visualization – Overview August 29, 2019 10 / 16

Projects

Lab projects – working with data

Analysis of the data using the methods discussed in the lecture will demonstrated through five projects scheduled throughout the course. The projects will require using statistical software. At the moment the most popular data mining softwares are R and Python, although commercial such as Matlab are also often utilized. R-package – free and very popular statistical package, very good for statistical computing although less compelling in handling large matrices and multivariate visualization We have opted for R-package to present analyses the data and the methods of data mining. The choice has been dictated by a large number of supporting materials in R that are available for illustration of data mining methods.

Data Mining and Visualization – Overview August 29, 2019 11 / 16

Projects

Downloading R-package and first steps

Statistical R-package available for free download here Available on any PC platform (Mac, Windows, Linux). Worry free and fast downloading procedure (a couple of minutes). We will be working in the command line window of R (most direct way of accessing R-package). No experience is required – all of the code that will be needed will be provided on our webpage! There some so-called R front-ends (such R Commander or R-Studio or Jupyter) that ease writing more complex programming in R

• nly a very basic R installation with the primitive

copy-and-paste-to-the-command-line approach is truly needed as a method of running the programs. You can do better if you wish so!

Data Mining and Visualization – Overview August 29, 2019 12 / 16

Projects

Example of a very simple R session

Suppose the following data are in file Table2 1.txt located at directory /Lecture1/Table2_1.txt

0.51 0.51 0.51 0.50 0.51 0.49 0.52 0.53 0.50 0.47 0.51 0.52 0.53 0.48 0.49 0.50 0.52 0.49 0.49 0.50 0.49 0.48 0.46 0.49 0.49 0.48 0.49 0.49 0.51 0.47 0.51 0.51 0.51 0.48 0.50 0.47 0.50 0.51 0.49 0.48 0.51 0.50 0.50 0.53 0.52 0.52 0.50 0.50 0.51 0.51

Then the following code reads the data and computing its mean and standard deviation:

#Getting data in a vector x=scan("/Data/Table2_1.txt") mean(x) #[1] 0.4998 sd(x) #[1] 0.01647385

# at the beginning of the line denotes a commentary (so the following characters are not interpreted by R when the lines are copied to the command line).

Data Mining and Visualization – Overview August 29, 2019 13 / 16

Projects

The same session through webpage

We go through steps of running this session using our webpage. Step One Create a directory in which you intend to work with your programs. Step Two Download data that you intend to work on to this directory. Step Three Download R-code the same directory and open the file by a convenient text editor of your choice. Step Four Open R and by copying and pasting run the programs from the code. Step Five Interpret the results.

Data Mining and Visualization – Overview August 29, 2019 14 / 16

Presentation

Final project/presentation – a study of a real data set

The final presentation will discuss a study of real data by means of learned data mining methods. Data Description – data, with a detailed description will be provided Problem Formulation – initial problem will be suggested in the description of the data, precise formulation will be required Analysis Methods – the choice of analytical tools and their discussion is expected Result Interpretation – analysis of the data has to be performed, results have to be presented and interpreted, final conclusions and suggestions need to be formulated

Data Mining and Visualization – Overview August 29, 2019 15 / 16

Presentation

Concluding thought We are drowning in information and starving for knowledge

Rutherford D. Roger∗

∗American librarian (Yale University)

Data Mining and Visualization – Overview August 29, 2019 16 / 16